Original article 235
Common and rare variants of the THBS1 gene associated with the risk for autism Lina Lua,c, Hui Guoa,c, Yu Penga,c, Guanglei Xunb,e, Yanling Liuc, Zhimin Xiongc, Di Tianc, Yalan Liuc, Wei Lic, Xiaojuan Xuc, Jingping Zhaob, Zhengmao Hua,c and Kun Xiaa,c,d Objectives Autism is a severe neurodevelopmental disorder. Many susceptible or causative genes have been identified, and most of them are related to synaptogenesis. The THBS1 gene encodes thrombospondin 1, which plays a critical role in synaptogenesis of the central nervous system in the developing brain. However, no study has been carried out revealing that THBS1 is an autism risk gene. Methods We analyzed the whole coding region and the 5′-untranslated region of the THBS1 gene in 313 autistic patients by Sanger sequencing, which was also used to analyze the identified variants in 350 normal controls. Association analysis was carried out using PLINK or R. Haplotype analysis was carried out using Haploview. Functional prediction and conservation analysis of missense variants were carried out using ANNOVAR. Results Twelve variants, including five common variants and seven rare variants, were identified in the THBS1 coding region and the 5′-untranslated region. Among them, one common variant (c.1567A > G:p.T523A) was significantly associated with autism (P < 0.05). Two rare variants (c.2429G > A:p.R810Q, c.3496G > C:p.E1166Q) were absent in the 350 controls and were not reported in the single
Introduction Autism is a severe neurodevelopmental disorder with impaired social interaction and language communication as well as restricted interests and/or stereotypic and repetitive activities as core clinical characterizations. In the last decade, the prevalence of autism has risen markedly. However, the etiology and pathogenesis of autism are still unclear. Twin and family studies provide the most convincing evidence for a genetic origin of autism (Beaudet, 2002). Recent genetic and genomic studies have identified a great number of candidate genes for autism (Abrahams and Geschwind, 2008), most of them, including but not limited to NRXN1, NLGN3, NLGN4, SHANK3, and CNTNAP2, are related to synapse formation, maturation, and/or modification after maturation. In addition, the onset of autism occurs before 3 years of age (Waterhouse et al., 1996), which is a crucial period for synapse formation and maturation (Südhof, 2008). These findings suggest that synaptic dysfunction may play an important role in the pathogenesis of autism. 0955-8829 © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins
nucleotide polymorphism database (dbSNP). Combined association analysis of the rare variants (minor allele frequency < 0.01) in patients and Asian samples in the 1000 genome project revealed a significant association between these rare variants and autism (P = 0.039). Conclusion Our data revealed that both common and rare variants of the THBS1 gene are associated with risk for autism, suggesting that THBS1 is a novel susceptible gene for autism. Psychiatr Genet 24:235–240 © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins. Psychiatric Genetics 2014, 24:235–240 Keywords: autism, mutation screening, synaptogenesis, THBS1 a
School of Life Sciences, bMental Health Institute of the Second Xiangya Hospital, cState Key Laboratory of Medical Genetics, dKey Laboratory of Medical Information Research, Central South University, Changsha, Hunan, China and e Mental Health Center of Shandong Province, Jinan, Shandong, China Correspondence to Kun Xia, PhD, State Key Laboratory of Medical Genetics, Central South University, 110 Xiangya Road, Changsha, Hunan, China Tel: + 86 731 8480 5357; fax: + 86 731 8447 8152; e-mail: [email protected] Received 10 October 2013 Revised 5 April 2014 Accepted 10 September 2014
Thrombospondin 1 (TSP1), encoded by the THBS1 gene, has been evidenced to play a critical role in synaptogenesis in the developing brain (Christopherson et al., 2005; Eroglu et al., 2009). TSP1 is involved in neuron development, migration, and synaptogenesis both in vivo and in vitro. Importantly, it could accelerate synaptogenesis through interaction with the extracellular domain of neuroligin 1, and exhibit effects similar to neurexin in inducing synaptogenesis (Xu et al., 2010). Neurexins are cell-surface molecules that bind neuroligins to form a heterophilic, Ca2 + -dependent complex at the central synapse. This complex is essential for efficient neurotransmission and is involved in the formation of synaptic contacts (Graf et al., 2004; Varoqueaux et al., 2006); furthermore, neurexins and neuroligins have been identified to be associated with the pathogenesis of autism by many studies. Thus, we hypothesize that the THBS1 gene is probably involved in the development of autism. To investigate the role of the THBS1 gene in autism, we analyzed the coding exons and the 5′-untranslated region DOI: 10.1097/YPG.0000000000000054
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
236 Psychiatric Genetics 2014, Vol 24 No 6
(UTR) of the THBS1 gene by Sanger sequencing. Our results show that both common and rare variants of the THBS1 gene are associated with autism.
Materials and methods Participants
Our study included 313 individuals (265 boys, 48 girls, the average age was 6 years, ranging from 3 to 18 years) with typical autism. All patients are from the Chinese Han population. Most patients were recruited from Yilin (Elim) Autism Training Department of the Qingdao Municipal Autism Research Institute, Qingdao, Shandong Province, China, and the rest were recruited from the Medical Psychological Research Center, the Second Xiangya Hospital of Central South University, Changsha, Hunan Province, China. Diagnosis was established by two experienced psychiatrists according to the Diagnostic and Statistical Manual of Mental Disorders, 4th ed. – text revision criteria for typical autism. Karyotyping analysis and the PCR procedure with denaturing polyacrylamide gel electrophoresis analysis were carried out for all patients to exclude those with chromosomal abnormalities or fragile X syndrome. Furthermore, patients were also excluded if they were identified with any other organic disease of the nervous system. The control group consisted of 350 healthy individuals (238 male, 112 female; the average age was 19 years, range 3–50 years), matched with the autism patients for sex and ethnicity. Written informed consent was obtained from all the patients and controls. This study was approved by the Human Ethical Committee of the State Key Lab of Medical Genetics and is compliant with the Code of Ethics of the World Medical Association (Dale and Salo, 1996). Mutation screening and statistical analysis
Genomic DNA was extracted from 5 ml peripheral blood samples of all individuals by standard proteinase K digestion and the phenol–chloroform method. The primer pairs spanning all exons, splicing sites, and 5′-UTRs were designed with the online Primer3 program (http://frodo.wi. mit.edu/). PCR amplification was performed using a thermocycler (Applied Biosystems Inc., Foster City, California, USA). The primer sequences and PCR conditions are available upon request. PCR products were sequenced using ABI PRISM 3730 DNA Sequencer (Applied Biosystems Inc.). Sanger sequencing was first performed in the 313 affected cases, the identified variants were further screened in the 350 unaffected controls. Association analysis was carried out using PLINK (Purcell et al., 2007) or R (http://www.r-project.org/). We first assessed the association between common variants and clinical outcome individually using the χ2-test or Fisher’s exact test under the allelic model. For the individually significant variants, we further carried out the association analysis under different models using the χ2-test or Fisher’s exact test (genotypic, trend, dominant, recessive)
and logistic regression (dominant, additive, recessive). Haplotype analysis was carried out using Haploview (Barrett et al., 2005). Combined association analysis for the rare variants is carried out using R. Functional prediction and conservation analysis of missense variants were carried out using ANNOVAR (Wang et al., 2010).
Results On sequencing the THBS1 gene in 313 autistic patients, 12 variants were identified. Among them, two variants were located in the 5′-UTR and 10 variants in the exons (Table 1). We identified four novel variants (c.2429G > A:p.R810Q, c.2876G > A:p.R959Q, c.3251G > A:p.G1084E, c.3496G > C:p.E1166Q) that are not reported in the single nucleotide polymorphism database (dbSNP)137, the 1000 genome project, and National Heart, Lung, and Blood Institute (NHLBI) exome sequencing project (ESP) exome sequencing data. Two novel rare variants (c.2429G > A:p.R810Q, c.3496G > C:p.E1166Q) were also absent in 350 normal controls. We further assessed these two variants in the parents; both were inherited from the parents, with no bias between maternal versus paternal transmission. To determine whether the common variants identified are associated with autism, we carried out association analysis using the χ2-test or Fisher’s exact test for the common variants individually under the allelic model. Two variants (c.1290G > A:p.K430K, c.1567A > G:p.T523A) showed a significant association with autism risk (Table 2). The variant c.1567A > G (p.T523A) is a missense SNP (rs2292305, minor allele frequency = 0.26). The P-value under the allelic model is 0.0004, with an odds ratio (OR) of 0.66, indicating a protective effect. The P-value remained significant even after correction for multiple testing (P = 0.002). The variant c.1290G > A (p.K430K) is a synonymous SNP (rs2229364, minor allele frequency = 0.35). The P-value under the allelic model is 0.026, with an OR of 0.77, also indicating a protective effect. The P-value did not survive the significant threshold after correction for multiple testing (P = 0.0634). However, haplotype analysis indicated that these two variants are in strong linkage disequilibrium (r2 = 0.89, D′ = 1; Fig. 1). To further assess the association between these two variants and autism, we also carried out the association analysis under other models (genotypic, trend, dominant, and recessive) using the χ2-test or Fisher’s exact test. All but the recessive model analysis indicated a significant association of these variants with autism (Table 3). We then carried out association analysis under three different genetic models (additive, dominant, and recessive) using logistic regression. The additive and dominant model, but not the recessive model, analyses indicated a significant association of these variants with autism (Table 4). These results show the association between these two variants and autism risk. To test whether autistic patients are burdened with the rare variants, we combined all the identified missense
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
THBS1 and autism Lu et al. 237
Table 1 Location
Variants identified in 5′-UTR and coding region of THBS1 Nucleotide change
Common variants 5′-UTR c. − 175G > A 5′-UTR c. − 138T > C Exon 9 c.1567A > G Exon 7 c.1290G > A Exon 8 c.1410C > T Rare variants Exon 2 c.455A > G Exon 11 c.1802A > G Exon 15 c.2429G > A Exon 15 c.2446G > A Exon 17 c.2876G > A Exon 18 c.3251G > A Exon 20 c.3496G > C
Amino acid change
Patients (AA/AB/BB) (n = 313)
MAF
Controls (AA/AB/BB) (n = 350)
MAF
snpID
NA NA p.T523A p.K430K p.N470N
25/148/140 25/149/139 19/144/150 22/140/151 21/141/151
0.3163 0.3179 0.2907 0.2939 0.2923
37/177/136 39/175/136 30/208/112 28/190/132 29/169/152
0.3586 0.3614 0.3829 0.3514 0.3243
rs1478605 rs1478604 rs2292305 rs2229364 rs2228261
p.Q152R p.N601S p.R810Q p.V816I p.R959Q p.G1084E p.E1166Q
0/2/311 0/6/307 0/2/311 0/14/299 0/1/312 0/3/310 0/1/312
0.0032 0.0096 0.0032 0.0224 0.0016 0.0048 0.0016
0/1/349 0/3/347 0/0/350 0/6/344 0/4/346 0/2/348 0/0/350
0.0014 0.0043 0.0000 0.0086 0.0057 0.0029 0.0000
rs183783284 rs201454661 Novel rs143790069 Novel Novel Novel
Position of variants is based on published cDNA sequence (NM_003246.2). AA/AB/BB, homozygous change/heterozygous change/normal; 5′-UTR, 5′-untranslated region.
Table 2
Association results under the allelic model for variants c.1290G > A:p.K430K and c.1567A > G:p.T523A
Variants
C_A
F_A
C_U
F_U
CHISQ
P
OR
L95
U95
Padj
c.1290G > A:p.K430K c.1567A > G:p.T523A
184 182
0.2939 0.2907
246 268
0.3514 0.3829
4.986 12.510
0.026 0.0004
0.768 0.661
0.609 0.525
0.969 0.832
0.0634 0.0020
Multiple testing was performed by Benjamini and Hochberg step-up false discovery rate control. C_A, number of changed alleles in affected individuals; C_U, number of changed alleles in unaffected individuals; F_A, frequency of changed alleles in affected individuals; F_U, frequency of changed alleles in unaffected individuals; L95, 95% lower confidence interval; U95, 95% upper confidence interval.
rare variants in the patients and 286 controls from the Asian population (CHB, CHS, JPT) in the 1000 Genome project and carried out association analysis. We observed a significant association between these rare variants and autism risk (P = 0.039, OR = 2.16).
Discussion The THBS1 gene (OMIM 188060), encoding TSP1, is located in the chromosome 15q14 region and consists of 21 exons. The domain architecture of human TSP1 and the location of the variants identified in this study are shown in Fig. 2. In this study, 12 variants were identified in the exons or 5′-UTRs of the THBS1 gene. Two common variants (p.K430K and p.T523A) showed significant association with autism risk (P < 0.05), and they are in strong disequilibrium (r2 = 0.89, D′ = 1). c.1290G > A:p.K430K is a common variant. The frequency of the changed allele A was significantly lower in patients with autism than in controls, and the OR is 0.77 under the allelic model, indicating that the c.G1290A variant within the THBS1 gene is protective against autism. Although it does not change the amino acid sequence, it may function as a silent mutation against autism. In addition, the nucleotide c.1290 is located at the 3′-end of exon 7 in the exon–intron junction. It may not just be a neutral change but a polymorphism that protects exon 7 from post-transcriptional skipping. However, further investigation is required to determine the function of this silent mutation involved in autism.
p.T523A is also a common variant, with a protective effect (OR = 0.66). The amino acid Thr523, which has a hydrophilic property, is substituted by alanine, which is hydrophobic. It is located in the third thrombospondin type 1 receptor (TSR3) domain. TSR domains are necessary for many cellular effects of TSP1. Importantly, TSR is also the site within TSP1 at which transforming growth factor-β (TGF-β) binds and is activated (Schultz-Cherry et al., 1994). TGF-β superfamily members have been shown to be widely expressed in developing and mature vertebrate nervous systems, and they play numerous roles in nervous system development (Lorentzon et al., 1996; Mehler et al., 1997; Withers et al., 2000). Recent studies have revealed that the TGF-β signal transduction pathway forms part of a feedback mechanism during synapse formation at the neuromuscular junction in Drosophila (Aberle et al., 2002; Marqués et al., 2002). Evidence for an involvement of TGF-β family members in synapse formation and plasticity has also been supported by studies on the marine mollusc Aplysia (Schuman, 1997; Zhang et al., 1997). Thus, the variant p.T523A identified in the TSR3 domain may play protective role in the process of activation of latent TGF-β by TSR domains, and consequent synapse formation. We also identified several rare variants (p.Q152R, p.N601S, p.R810Q, p.V816I, p.R959Q, p.G1084E, p.E1166Q) in the coding region of the THBS1 gene in autistic patients. A significant association was identified between these rare variants and autism risk (P = 0.039,
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
238 Psychiatric Genetics 2014, Vol 24 No 6
2 Table 3 Association results using the χ -test or Fisher’s exact test for the two significant variants under different models
Fig. 1
Model
chr15 37 668 000 Genotyped s A G
C T
AFF
UNAFF
c.1290G > A:p.K430K:rs2229364 GENO 22/140/151 28/190/132 TREND 184/442 246/454 DOM 162/151 218/132 REC 22/291 28/322 c.1567A > G:p.T523A:rs2292305 GENO 19/144/150 30/208/112 TREND 182/444 268/432 DOM 163/150 238/112 REC 19/294 30/320
χ2
d.f.
P
7.53 5.769 7.487 0.2235
2 1 1 1
0.02317 0.01631 0.006216 0.6364
2 1 1 1
0.0001502 9.02E-05 2.83E-05 0.2191
17.61 15.33 17.53 1.51
ADOM, dominant model; AFF, affected; GENO, genotypic model; REC, recessive model; TREND, trend model; UNAFF, unaffected.
Entrez gene NM_003246
Table 4 Association results using logistic regression results of the two significant variants under different models
THBS1 : thromb
rs2292305
rs2228261
Model
Block 1 (0 kb) 1
2
89
Haplotype analysis for the two common variants (c.1290G > A:p.K430 and c.1567A > G:p.T523A) associated with autism risk. The two variants are in strong linkage disequilibrium (r2 = 0.89, D′ = 1).
OR = 2.16), and most of these variants are predicted to be damaging or possibly damaging by SIFT or PolyPhen2 and conserved by GREE + + (Table 5). All of them except p.Q152R are located in the important C-terminal region, which folds and functions as a single structural unit through multiple intracellular interactions between the EGF-like domain, the type 3 repeats, and the L-lection domain (Adams and Lawler, 2011). Importantly, Arg810 and Val816, conserved across different species, are located in the fourth repeat of type 3 repeats, which bind to calcium through the 12 interacting Ca2 + -binding loops in this region (Lawler and Simons, 1983; Lawler and Hynes, 1986; Sun et al., 1992). Both Arg810 and Val816
NMISS
STAT
c.1290G > A:p.K430K:rs2229364 ADD 663 − 2.39 DOM 663 − 2.73 REC 663 − 0.47 c.1567A > G:p.T523A:rs2292305 ADD 663 − 3.88 DOM 663 − 4.17 REC 663 − 1.22
P
OR
L95
U95
0.01667 0.006323 0.6366
0.74 0.65 0.87
0.57 0.48 0.49
0.95 0.89 1.55
0.0001041 3.12E-05 0.2212
0.60 0.51 0.69
0.46 0.37 0.38
0.78 0.70 1.25
ADD, additive model; DOM, dominant model; L95, 95% lower confidence interval; NMISS, number of missing; OR, odds ratio; REC, recessive model; U95, 95% upper confidence interval.
are located in the fifth loop (Sun et al., 1992); therefore, the variants p.R810Q and p.V816I may affect the interactions of the fifth loop with other loops, further affecting the ability of type 3 repeats to bind to calcium, which plays an important role in the functional activities with a conformational change in the TSP molecules (Lawler and Simons, 1983). This conformational change in TSP1 raises the possibility that the variant p.V816I may affect more aspects of TSP1 function. Furthermore, TSP1, as an adhesive extracellular matrix (ECM) molecule, is necessary and sufficient for synapse formation (Washbourne et al., 2004), and this function of TSP1 depends on its presence in the ECM. Interestingly, the retention of TSP1 within the ECM was mediated by the highly conserved C-terminal region in the trimeric form (Adams et al., 2008). Thus, these rare variants may alter the folding of the C-terminal region, thus affecting the retention of TSP1 within the ECM, further affecting synapse formation. Conclusively, we identified two common variants in the THBS1 gene that were significantly associated with the risk for autism individually, and the combined rare variants also present a significant association with autism risk. Our study strongly suggests that the THBS1 gene is a novel susceptible gene for autism, and further study should be conducted to explore the association between THBS1 and autism from both genetic and functional perspectives.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
THBS1 and autism Lu et al. 239
Fig. 2
(a)
LG
Coil
vWF_C
p.N601S
p.K430K
p.Q152R
Type 3
EGF
TSR
L-lectin
p.R959Q p.G1084E p.E1166Q
p.R810Q p.V816I
p.N470N p.T523A (b) Human Chimp Gorilla Orangutan Rhesus Baboon Marmoset Elephant Cow Horse Cat Dog Guinea pig Mouse Shrew
... TGQWK ... CDKRF ... QMNGK ... NPTPQ ... CFNHN ... NERDN ... QYVYN ... FRRFQ ... HTGNT ... KYECR ... ... TGQWK ... CDKRF ... QMNGK ... NPTPQ ... CFNHN ... NERDN ... QYVYN ... FRRFQ ... HTGNT ... KYECR ... ... TGQWK ... CDKRF ... QMNGK ... NPTPQ ... CFNHN ... NERDN ... QYVYN ... FRRFQ ... HTGNT ... KYECR ... ... TGQWK ... CDKRF ... QMNGK ... NPTPQ ... CFNHN ... NERDN ... QYVYN ... FRRFQ ... HTGNT ... KYECR ... ... TGQWK ... CDKRF ... QMNGK ... NPTPQ ... CFNHN ... NERDN ... QYVYN ... FRRFQ ... HTGNT ... KYECR ... ... TGQWK ... CDKRF ... QMNGK ... NPTPQ ... CFNHN ... NERDN ... QYVYN ... FRRFQ ... HTGNT ... KYECR ... ... TGQWK ... CDKRF ... QMNGK ... NPAPQ ... CFNHN ... NERDN ... QYVYN ... FRRFQ ... HTGNT ... KYECR ... ... TGQWK ... CDKRF ... QMNGK ... NPTPQ ... CFNHN ... NERDN ... QYVYN ... FRRFQ ... HTGNT ... KYECR ... ... TGQWK ... CDKRF ... QMNGK ... NPTPQ ... CFNHN ... NERDN ... QYVYN ... FRRFQ ... HTGNT ... KYECR ... ... TGQWK ... CDKRF ... QMNGK ... NPTPQ ... CFNHN ... NERDN ... QYVYN ... FRRFQ ... HTGNT ... KYECR ... ... TGQWK ... CDKRF ... QMNGK ... NPTPQ ... CFNHN ... NERDN ... QYVYN ... FRRFQ ... HTGNT ... KYECR ... ... TGQWK ... CDKRF ... QMNGK ... NPTPQ ... CFNHN ... NERDN ... QYVYN ... FRRFQ ... HTGNT ... KYECR ... ... TGQWK ... CDKRF ... QMNGK ... NPMPQ ... CFNHN ... NERDN ... QYVYN ... FRRFQ ... HTGNT ... KYECR ... ... TGQWK ... CDKRF ... QMNGK ... NPTPQ ... CFNHN ... NERDN ... QYVYN ... FRQFQ ... HTGNT ... KYECR ... ... TGQWK ... CDKRF ... QMNGK ... NPPPQ ... CFNHN ... NERDN ... RYVYN ... FRRFQ ... HTGNT ... KYECR ... p.Q152R
p.K430K
p.N470N
p.T523A p.N601S p.R810Q p.V816I p.R959Q p.G1084E p.E1166Q
(a) Domain architecture of human thrombospondin 1 (TSP1), and variants identified in different domains. This architecture was drawn according to the data on human TSP1 (P07996) in the website http://web.expasy.org/docs/swiss-prot_guideline.html. Locations of the variants in the deferent domain are shown with black arrows. (b) Alignment of the TSP1 protein sequence with homologous protein sequences of other species. Most variants were completely conserved in different species. Amino acids that were not conserved are shaded in gray. EGF, epidermal growth factor-like domains; LG, laminin G-like amino-terminal domain; L-lectin, L-type lectin-like domain; TSR, thrombospondin type 1 domains; Type 3, thrombospondin type 3 repeats; vWF_C, von Willebrand factor type C domain.
Table 5
Functional prediction and conservation analysis of nonsynonymous variants using SIFT, PolyPhen2, and GERP + +
Variant c.455A > G:p.Q152R c.1567A > G:p.T523A c.1802A > G:p.N601S c.2429G > A:p.R810Q c.2446G > A:p.V816I c.2876G > A:p.R959Q c.3251G > A:p.G1084E c.3496G > C:p.E1166Q
dbSNP
SIFT_score
SIFT_pred
PolyPhen2_score
PolyPhen2_pred
GERP + +
rs183783284 rs2292305 rs201454661 NA rs143790069 NA NA NA
0.97 0.18 0.17 0.69 0.92 0.39 1 0.87
D T T T T T D T
0.028 0 0 0.179 0.041 0.334 1 0.988
B B B P B P D D
4.53 −1.26 4.2 5.35 5.35 4.34 5.43 5.39
B, benign; D, damaging; dbSNP, single nucleotide polymorphism database; P, possibly damaging, T, tolerant.
Acknowledgements The authors thank all the families that participated and collaborated with us in this study. They also thank the Yilin (Elim) Autism Training Department of the Qingdao Municipal Autism Research Institute, which provided a lot of help in the collecting samples. They also thank Professor Jiada Li, who assisted with proofreading of the manuscript. This work was supported by the National Basic Research Program of China (2012CB517902) and
the National Natural Science Foundation of China (81330027, 81161120544 and 81301172). Correspondence should be addressed to Kun Xia ([email protected]) or Zhengmao Hu (huzhengmao@ sklmg.edu.cn). Conflicts of interest
There are no conflicts of interest.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
240 Psychiatric Genetics 2014, Vol 24 No 6
References Aberle H, Haghighi AP, Fetter RD, McCabe BD, Magalhães TR, Goodman CS (2002). Wishful thinking encodes a BMP type II receptor that regulates synaptic growth in Drosophila. Neuron 33:545–558. Abrahams BS, Geschwind DH (2008). Advances in autism genetics: on the threshold of a new neurobiology. Nat Rev Genet 9:341–355. Adams JC, Lawler J (2011). The thrombospondins. Cold Spring Harb Perspect Biol 3:a009712. Adams JC, Bentley AA, Kvansakul M, Hatherley D, Hohenester E (2008). Extracellular matrix retention of thrombospondin 1 is controlled by its conserved C-terminal region. J Cell Sci 121:784–795. Barrett JC, Fry B, Maller J, Daly MJ (2005). Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics 21:263–265. Beaudet AL (2002). Is medical genetics neglecting epigenetics? Genet Med 4:399–402. Christopherson KS, Ullian EM, Stokes CC, Mullowney CE, Hell JW, Agah A, et al. (2005). Thrombospondins are astrocyte-secreted proteins that promote CNS synaptogenesis. Cell 120:421–433. Dale O, Salo M (1996). The Helsinki Declaration, research guidelines and regulations: present and future editorial aspects. Acta Anaesthesiol Scand 40:771–772. Eroglu C, Allen NJ, Susman MW, O’Rourke NA, Park CY, Ozkan E, et al. (2009). Gabapentin receptor alpha2delta-1 is a neuronal thrombospondin receptor responsible for excitatory CNS synaptogenesis. Cell 139:380–392. Graf ER, Zhang X, Jin SX, Linhoff MW, Craig AM (2004). Neurexins induce differentiation of GABA and glutamate postsynaptic specializations via neuroligins. Cell 119:1013–1026. Lawler J, Simons ER (1983). Cooperative binding of calcium to thrombospondin. The effect of calcium on the circular dichroism and limited tryptic digestion of thrombospondin. J Biol Chem 258:12098–12101. Lawler J, Hynes RO (1986). The structure of human thrombospondin, an adhesive glycoprotein with multiple calcium-binding sites and homologies with several different proteins. J Cell Biol 103:1635–1648. Lorentzon M, Hoffer B, Ebendal T, Olson L, Tomac A (1996). Habrec1, a novel serine/threonine kinase TGF-beta type I-like receptor, has a specific cellular expression suggesting function in the developing organism and adult brain. Exp Neurol 142:351–360.
Marqués G, Bao H, Haerry TE, Shimell MJ, Duchek P, Zhang B, O’Connor MB (2002). The Drosophila BMP type II receptor Wishful Thinking regulates neuromuscular synapse morphology and function. Neuron 33: 529–543. Mehler MF, Mabie PC, Zhang D, Kessler JA (1997). Bone morphogenetic proteins in the nervous system. Trends Neurosci 20:309–317. Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D, et al. (2007). PLINK: a tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet 81:559–575. Schultz-Cherry S, Lawler J, Murphy-Ullrich JE (1994). The type 1 repeats of thrombospondin 1 activate latent transforming growth factor-beta. J Biol Chem 269:26783–26788. Schuman EM (1997). Synapse specificity and long-term information storage. Neuron 18:339–342. Südhof TC (2008). Neuroligins and neurexins link synaptic function to cognitive disease. Nature 455:903–911. Sun X, Skorstengaard K, Mosher DF (1992). Disulfides modulate RGD-inhibitable cell adhesive activity of thrombospondin. J Cell Biol 118:693–701. Varoqueaux F, Aramuni G, Rawson RL, Mohrmann R, Missler M, Gottmann K, et al. (2006). Neuroligins determine synapse maturation and function. Neuron 51:741–754. Wang K, Li M, Hakonarson H (2010). ANNOVAR: functional annotation of genetic variants from high-throughput sequencing data. Nucleic Acids Res 38:e164. Washbourne P, Dityatev A, Scheiffele P, Biederer T, Weiner JA, Christopherson KS, El-Husseini A (2004). Cell adhesion molecules in synapse formation. J Neurosci 24:9244–9249. Waterhouse L, Morris R, Allen D, Dunn M, Fein D, Feinstein C, et al. (1996). Diagnosis and classification in autism. J Autism Dev Disord 26:59–86. Withers GS, Higgins D, Charette M, Banker G (2000). Bone morphogenetic protein-7 enhances dendritic growth and receptivity to innervation in cultured hippocampal neurons. Eur J Neurosci 12:106–116. Xu J, Xiao N, Xia J (2010). Thrombospondin 1 accelerates synaptogenesis in hippocampal neurons through neuroligin 1. Nat Neurosci 13:22–24. Zhang F, Endo S, Cleary LJ, Eskin A, Byrne JH (1997). Role of transforming growth factor-beta in long-term synaptic facilitation in Aplysia. Science 275:1318–1320.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Common and rare variants of the THBS1 gene associated with the risk for autism Lina Lua,c, Hui Guoa,c, Yu Penga,c, Guanglei Xunb,e, Yanling Liuc, Zhimin Xiongc, Di Tianc, Yalan Liuc, Wei Lic, Xiaojuan Xuc, Jingping Zhaob, Zhengmao Hua,c and Kun Xiaa,c,d Objectives Autism is a severe neurodevelopmental disorder. Many susceptible or causative genes have been identified, and most of them are related to synaptogenesis. The THBS1 gene encodes thrombospondin 1, which plays a critical role in synaptogenesis of the central nervous system in the developing brain. However, no study has been carried out revealing that THBS1 is an autism risk gene. Methods We analyzed the whole coding region and the 5′-untranslated region of the THBS1 gene in 313 autistic patients by Sanger sequencing, which was also used to analyze the identified variants in 350 normal controls. Association analysis was carried out using PLINK or R. Haplotype analysis was carried out using Haploview. Functional prediction and conservation analysis of missense variants were carried out using ANNOVAR. Results Twelve variants, including five common variants and seven rare variants, were identified in the THBS1 coding region and the 5′-untranslated region. Among them, one common variant (c.1567A > G:p.T523A) was significantly associated with autism (P < 0.05). Two rare variants (c.2429G > A:p.R810Q, c.3496G > C:p.E1166Q) were absent in the 350 controls and were not reported in the single
Introduction Autism is a severe neurodevelopmental disorder with impaired social interaction and language communication as well as restricted interests and/or stereotypic and repetitive activities as core clinical characterizations. In the last decade, the prevalence of autism has risen markedly. However, the etiology and pathogenesis of autism are still unclear. Twin and family studies provide the most convincing evidence for a genetic origin of autism (Beaudet, 2002). Recent genetic and genomic studies have identified a great number of candidate genes for autism (Abrahams and Geschwind, 2008), most of them, including but not limited to NRXN1, NLGN3, NLGN4, SHANK3, and CNTNAP2, are related to synapse formation, maturation, and/or modification after maturation. In addition, the onset of autism occurs before 3 years of age (Waterhouse et al., 1996), which is a crucial period for synapse formation and maturation (Südhof, 2008). These findings suggest that synaptic dysfunction may play an important role in the pathogenesis of autism. 0955-8829 © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins
nucleotide polymorphism database (dbSNP). Combined association analysis of the rare variants (minor allele frequency < 0.01) in patients and Asian samples in the 1000 genome project revealed a significant association between these rare variants and autism (P = 0.039). Conclusion Our data revealed that both common and rare variants of the THBS1 gene are associated with risk for autism, suggesting that THBS1 is a novel susceptible gene for autism. Psychiatr Genet 24:235–240 © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins. Psychiatric Genetics 2014, 24:235–240 Keywords: autism, mutation screening, synaptogenesis, THBS1 a
School of Life Sciences, bMental Health Institute of the Second Xiangya Hospital, cState Key Laboratory of Medical Genetics, dKey Laboratory of Medical Information Research, Central South University, Changsha, Hunan, China and e Mental Health Center of Shandong Province, Jinan, Shandong, China Correspondence to Kun Xia, PhD, State Key Laboratory of Medical Genetics, Central South University, 110 Xiangya Road, Changsha, Hunan, China Tel: + 86 731 8480 5357; fax: + 86 731 8447 8152; e-mail: [email protected] Received 10 October 2013 Revised 5 April 2014 Accepted 10 September 2014
Thrombospondin 1 (TSP1), encoded by the THBS1 gene, has been evidenced to play a critical role in synaptogenesis in the developing brain (Christopherson et al., 2005; Eroglu et al., 2009). TSP1 is involved in neuron development, migration, and synaptogenesis both in vivo and in vitro. Importantly, it could accelerate synaptogenesis through interaction with the extracellular domain of neuroligin 1, and exhibit effects similar to neurexin in inducing synaptogenesis (Xu et al., 2010). Neurexins are cell-surface molecules that bind neuroligins to form a heterophilic, Ca2 + -dependent complex at the central synapse. This complex is essential for efficient neurotransmission and is involved in the formation of synaptic contacts (Graf et al., 2004; Varoqueaux et al., 2006); furthermore, neurexins and neuroligins have been identified to be associated with the pathogenesis of autism by many studies. Thus, we hypothesize that the THBS1 gene is probably involved in the development of autism. To investigate the role of the THBS1 gene in autism, we analyzed the coding exons and the 5′-untranslated region DOI: 10.1097/YPG.0000000000000054
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
236 Psychiatric Genetics 2014, Vol 24 No 6
(UTR) of the THBS1 gene by Sanger sequencing. Our results show that both common and rare variants of the THBS1 gene are associated with autism.
Materials and methods Participants
Our study included 313 individuals (265 boys, 48 girls, the average age was 6 years, ranging from 3 to 18 years) with typical autism. All patients are from the Chinese Han population. Most patients were recruited from Yilin (Elim) Autism Training Department of the Qingdao Municipal Autism Research Institute, Qingdao, Shandong Province, China, and the rest were recruited from the Medical Psychological Research Center, the Second Xiangya Hospital of Central South University, Changsha, Hunan Province, China. Diagnosis was established by two experienced psychiatrists according to the Diagnostic and Statistical Manual of Mental Disorders, 4th ed. – text revision criteria for typical autism. Karyotyping analysis and the PCR procedure with denaturing polyacrylamide gel electrophoresis analysis were carried out for all patients to exclude those with chromosomal abnormalities or fragile X syndrome. Furthermore, patients were also excluded if they were identified with any other organic disease of the nervous system. The control group consisted of 350 healthy individuals (238 male, 112 female; the average age was 19 years, range 3–50 years), matched with the autism patients for sex and ethnicity. Written informed consent was obtained from all the patients and controls. This study was approved by the Human Ethical Committee of the State Key Lab of Medical Genetics and is compliant with the Code of Ethics of the World Medical Association (Dale and Salo, 1996). Mutation screening and statistical analysis
Genomic DNA was extracted from 5 ml peripheral blood samples of all individuals by standard proteinase K digestion and the phenol–chloroform method. The primer pairs spanning all exons, splicing sites, and 5′-UTRs were designed with the online Primer3 program (http://frodo.wi. mit.edu/). PCR amplification was performed using a thermocycler (Applied Biosystems Inc., Foster City, California, USA). The primer sequences and PCR conditions are available upon request. PCR products were sequenced using ABI PRISM 3730 DNA Sequencer (Applied Biosystems Inc.). Sanger sequencing was first performed in the 313 affected cases, the identified variants were further screened in the 350 unaffected controls. Association analysis was carried out using PLINK (Purcell et al., 2007) or R (http://www.r-project.org/). We first assessed the association between common variants and clinical outcome individually using the χ2-test or Fisher’s exact test under the allelic model. For the individually significant variants, we further carried out the association analysis under different models using the χ2-test or Fisher’s exact test (genotypic, trend, dominant, recessive)
and logistic regression (dominant, additive, recessive). Haplotype analysis was carried out using Haploview (Barrett et al., 2005). Combined association analysis for the rare variants is carried out using R. Functional prediction and conservation analysis of missense variants were carried out using ANNOVAR (Wang et al., 2010).
Results On sequencing the THBS1 gene in 313 autistic patients, 12 variants were identified. Among them, two variants were located in the 5′-UTR and 10 variants in the exons (Table 1). We identified four novel variants (c.2429G > A:p.R810Q, c.2876G > A:p.R959Q, c.3251G > A:p.G1084E, c.3496G > C:p.E1166Q) that are not reported in the single nucleotide polymorphism database (dbSNP)137, the 1000 genome project, and National Heart, Lung, and Blood Institute (NHLBI) exome sequencing project (ESP) exome sequencing data. Two novel rare variants (c.2429G > A:p.R810Q, c.3496G > C:p.E1166Q) were also absent in 350 normal controls. We further assessed these two variants in the parents; both were inherited from the parents, with no bias between maternal versus paternal transmission. To determine whether the common variants identified are associated with autism, we carried out association analysis using the χ2-test or Fisher’s exact test for the common variants individually under the allelic model. Two variants (c.1290G > A:p.K430K, c.1567A > G:p.T523A) showed a significant association with autism risk (Table 2). The variant c.1567A > G (p.T523A) is a missense SNP (rs2292305, minor allele frequency = 0.26). The P-value under the allelic model is 0.0004, with an odds ratio (OR) of 0.66, indicating a protective effect. The P-value remained significant even after correction for multiple testing (P = 0.002). The variant c.1290G > A (p.K430K) is a synonymous SNP (rs2229364, minor allele frequency = 0.35). The P-value under the allelic model is 0.026, with an OR of 0.77, also indicating a protective effect. The P-value did not survive the significant threshold after correction for multiple testing (P = 0.0634). However, haplotype analysis indicated that these two variants are in strong linkage disequilibrium (r2 = 0.89, D′ = 1; Fig. 1). To further assess the association between these two variants and autism, we also carried out the association analysis under other models (genotypic, trend, dominant, and recessive) using the χ2-test or Fisher’s exact test. All but the recessive model analysis indicated a significant association of these variants with autism (Table 3). We then carried out association analysis under three different genetic models (additive, dominant, and recessive) using logistic regression. The additive and dominant model, but not the recessive model, analyses indicated a significant association of these variants with autism (Table 4). These results show the association between these two variants and autism risk. To test whether autistic patients are burdened with the rare variants, we combined all the identified missense
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
THBS1 and autism Lu et al. 237
Table 1 Location
Variants identified in 5′-UTR and coding region of THBS1 Nucleotide change
Common variants 5′-UTR c. − 175G > A 5′-UTR c. − 138T > C Exon 9 c.1567A > G Exon 7 c.1290G > A Exon 8 c.1410C > T Rare variants Exon 2 c.455A > G Exon 11 c.1802A > G Exon 15 c.2429G > A Exon 15 c.2446G > A Exon 17 c.2876G > A Exon 18 c.3251G > A Exon 20 c.3496G > C
Amino acid change
Patients (AA/AB/BB) (n = 313)
MAF
Controls (AA/AB/BB) (n = 350)
MAF
snpID
NA NA p.T523A p.K430K p.N470N
25/148/140 25/149/139 19/144/150 22/140/151 21/141/151
0.3163 0.3179 0.2907 0.2939 0.2923
37/177/136 39/175/136 30/208/112 28/190/132 29/169/152
0.3586 0.3614 0.3829 0.3514 0.3243
rs1478605 rs1478604 rs2292305 rs2229364 rs2228261
p.Q152R p.N601S p.R810Q p.V816I p.R959Q p.G1084E p.E1166Q
0/2/311 0/6/307 0/2/311 0/14/299 0/1/312 0/3/310 0/1/312
0.0032 0.0096 0.0032 0.0224 0.0016 0.0048 0.0016
0/1/349 0/3/347 0/0/350 0/6/344 0/4/346 0/2/348 0/0/350
0.0014 0.0043 0.0000 0.0086 0.0057 0.0029 0.0000
rs183783284 rs201454661 Novel rs143790069 Novel Novel Novel
Position of variants is based on published cDNA sequence (NM_003246.2). AA/AB/BB, homozygous change/heterozygous change/normal; 5′-UTR, 5′-untranslated region.
Table 2
Association results under the allelic model for variants c.1290G > A:p.K430K and c.1567A > G:p.T523A
Variants
C_A
F_A
C_U
F_U
CHISQ
P
OR
L95
U95
Padj
c.1290G > A:p.K430K c.1567A > G:p.T523A
184 182
0.2939 0.2907
246 268
0.3514 0.3829
4.986 12.510
0.026 0.0004
0.768 0.661
0.609 0.525
0.969 0.832
0.0634 0.0020
Multiple testing was performed by Benjamini and Hochberg step-up false discovery rate control. C_A, number of changed alleles in affected individuals; C_U, number of changed alleles in unaffected individuals; F_A, frequency of changed alleles in affected individuals; F_U, frequency of changed alleles in unaffected individuals; L95, 95% lower confidence interval; U95, 95% upper confidence interval.
rare variants in the patients and 286 controls from the Asian population (CHB, CHS, JPT) in the 1000 Genome project and carried out association analysis. We observed a significant association between these rare variants and autism risk (P = 0.039, OR = 2.16).
Discussion The THBS1 gene (OMIM 188060), encoding TSP1, is located in the chromosome 15q14 region and consists of 21 exons. The domain architecture of human TSP1 and the location of the variants identified in this study are shown in Fig. 2. In this study, 12 variants were identified in the exons or 5′-UTRs of the THBS1 gene. Two common variants (p.K430K and p.T523A) showed significant association with autism risk (P < 0.05), and they are in strong disequilibrium (r2 = 0.89, D′ = 1). c.1290G > A:p.K430K is a common variant. The frequency of the changed allele A was significantly lower in patients with autism than in controls, and the OR is 0.77 under the allelic model, indicating that the c.G1290A variant within the THBS1 gene is protective against autism. Although it does not change the amino acid sequence, it may function as a silent mutation against autism. In addition, the nucleotide c.1290 is located at the 3′-end of exon 7 in the exon–intron junction. It may not just be a neutral change but a polymorphism that protects exon 7 from post-transcriptional skipping. However, further investigation is required to determine the function of this silent mutation involved in autism.
p.T523A is also a common variant, with a protective effect (OR = 0.66). The amino acid Thr523, which has a hydrophilic property, is substituted by alanine, which is hydrophobic. It is located in the third thrombospondin type 1 receptor (TSR3) domain. TSR domains are necessary for many cellular effects of TSP1. Importantly, TSR is also the site within TSP1 at which transforming growth factor-β (TGF-β) binds and is activated (Schultz-Cherry et al., 1994). TGF-β superfamily members have been shown to be widely expressed in developing and mature vertebrate nervous systems, and they play numerous roles in nervous system development (Lorentzon et al., 1996; Mehler et al., 1997; Withers et al., 2000). Recent studies have revealed that the TGF-β signal transduction pathway forms part of a feedback mechanism during synapse formation at the neuromuscular junction in Drosophila (Aberle et al., 2002; Marqués et al., 2002). Evidence for an involvement of TGF-β family members in synapse formation and plasticity has also been supported by studies on the marine mollusc Aplysia (Schuman, 1997; Zhang et al., 1997). Thus, the variant p.T523A identified in the TSR3 domain may play protective role in the process of activation of latent TGF-β by TSR domains, and consequent synapse formation. We also identified several rare variants (p.Q152R, p.N601S, p.R810Q, p.V816I, p.R959Q, p.G1084E, p.E1166Q) in the coding region of the THBS1 gene in autistic patients. A significant association was identified between these rare variants and autism risk (P = 0.039,
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
238 Psychiatric Genetics 2014, Vol 24 No 6
2 Table 3 Association results using the χ -test or Fisher’s exact test for the two significant variants under different models
Fig. 1
Model
chr15 37 668 000 Genotyped s A G
C T
AFF
UNAFF
c.1290G > A:p.K430K:rs2229364 GENO 22/140/151 28/190/132 TREND 184/442 246/454 DOM 162/151 218/132 REC 22/291 28/322 c.1567A > G:p.T523A:rs2292305 GENO 19/144/150 30/208/112 TREND 182/444 268/432 DOM 163/150 238/112 REC 19/294 30/320
χ2
d.f.
P
7.53 5.769 7.487 0.2235
2 1 1 1
0.02317 0.01631 0.006216 0.6364
2 1 1 1
0.0001502 9.02E-05 2.83E-05 0.2191
17.61 15.33 17.53 1.51
ADOM, dominant model; AFF, affected; GENO, genotypic model; REC, recessive model; TREND, trend model; UNAFF, unaffected.
Entrez gene NM_003246
Table 4 Association results using logistic regression results of the two significant variants under different models
THBS1 : thromb
rs2292305
rs2228261
Model
Block 1 (0 kb) 1
2
89
Haplotype analysis for the two common variants (c.1290G > A:p.K430 and c.1567A > G:p.T523A) associated with autism risk. The two variants are in strong linkage disequilibrium (r2 = 0.89, D′ = 1).
OR = 2.16), and most of these variants are predicted to be damaging or possibly damaging by SIFT or PolyPhen2 and conserved by GREE + + (Table 5). All of them except p.Q152R are located in the important C-terminal region, which folds and functions as a single structural unit through multiple intracellular interactions between the EGF-like domain, the type 3 repeats, and the L-lection domain (Adams and Lawler, 2011). Importantly, Arg810 and Val816, conserved across different species, are located in the fourth repeat of type 3 repeats, which bind to calcium through the 12 interacting Ca2 + -binding loops in this region (Lawler and Simons, 1983; Lawler and Hynes, 1986; Sun et al., 1992). Both Arg810 and Val816
NMISS
STAT
c.1290G > A:p.K430K:rs2229364 ADD 663 − 2.39 DOM 663 − 2.73 REC 663 − 0.47 c.1567A > G:p.T523A:rs2292305 ADD 663 − 3.88 DOM 663 − 4.17 REC 663 − 1.22
P
OR
L95
U95
0.01667 0.006323 0.6366
0.74 0.65 0.87
0.57 0.48 0.49
0.95 0.89 1.55
0.0001041 3.12E-05 0.2212
0.60 0.51 0.69
0.46 0.37 0.38
0.78 0.70 1.25
ADD, additive model; DOM, dominant model; L95, 95% lower confidence interval; NMISS, number of missing; OR, odds ratio; REC, recessive model; U95, 95% upper confidence interval.
are located in the fifth loop (Sun et al., 1992); therefore, the variants p.R810Q and p.V816I may affect the interactions of the fifth loop with other loops, further affecting the ability of type 3 repeats to bind to calcium, which plays an important role in the functional activities with a conformational change in the TSP molecules (Lawler and Simons, 1983). This conformational change in TSP1 raises the possibility that the variant p.V816I may affect more aspects of TSP1 function. Furthermore, TSP1, as an adhesive extracellular matrix (ECM) molecule, is necessary and sufficient for synapse formation (Washbourne et al., 2004), and this function of TSP1 depends on its presence in the ECM. Interestingly, the retention of TSP1 within the ECM was mediated by the highly conserved C-terminal region in the trimeric form (Adams et al., 2008). Thus, these rare variants may alter the folding of the C-terminal region, thus affecting the retention of TSP1 within the ECM, further affecting synapse formation. Conclusively, we identified two common variants in the THBS1 gene that were significantly associated with the risk for autism individually, and the combined rare variants also present a significant association with autism risk. Our study strongly suggests that the THBS1 gene is a novel susceptible gene for autism, and further study should be conducted to explore the association between THBS1 and autism from both genetic and functional perspectives.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
THBS1 and autism Lu et al. 239
Fig. 2
(a)
LG
Coil
vWF_C
p.N601S
p.K430K
p.Q152R
Type 3
EGF
TSR
L-lectin
p.R959Q p.G1084E p.E1166Q
p.R810Q p.V816I
p.N470N p.T523A (b) Human Chimp Gorilla Orangutan Rhesus Baboon Marmoset Elephant Cow Horse Cat Dog Guinea pig Mouse Shrew
... TGQWK ... CDKRF ... QMNGK ... NPTPQ ... CFNHN ... NERDN ... QYVYN ... FRRFQ ... HTGNT ... KYECR ... ... TGQWK ... CDKRF ... QMNGK ... NPTPQ ... CFNHN ... NERDN ... QYVYN ... FRRFQ ... HTGNT ... KYECR ... ... TGQWK ... CDKRF ... QMNGK ... NPTPQ ... CFNHN ... NERDN ... QYVYN ... FRRFQ ... HTGNT ... KYECR ... ... TGQWK ... CDKRF ... QMNGK ... NPTPQ ... CFNHN ... NERDN ... QYVYN ... FRRFQ ... HTGNT ... KYECR ... ... TGQWK ... CDKRF ... QMNGK ... NPTPQ ... CFNHN ... NERDN ... QYVYN ... FRRFQ ... HTGNT ... KYECR ... ... TGQWK ... CDKRF ... QMNGK ... NPTPQ ... CFNHN ... NERDN ... QYVYN ... FRRFQ ... HTGNT ... KYECR ... ... TGQWK ... CDKRF ... QMNGK ... NPAPQ ... CFNHN ... NERDN ... QYVYN ... FRRFQ ... HTGNT ... KYECR ... ... TGQWK ... CDKRF ... QMNGK ... NPTPQ ... CFNHN ... NERDN ... QYVYN ... FRRFQ ... HTGNT ... KYECR ... ... TGQWK ... CDKRF ... QMNGK ... NPTPQ ... CFNHN ... NERDN ... QYVYN ... FRRFQ ... HTGNT ... KYECR ... ... TGQWK ... CDKRF ... QMNGK ... NPTPQ ... CFNHN ... NERDN ... QYVYN ... FRRFQ ... HTGNT ... KYECR ... ... TGQWK ... CDKRF ... QMNGK ... NPTPQ ... CFNHN ... NERDN ... QYVYN ... FRRFQ ... HTGNT ... KYECR ... ... TGQWK ... CDKRF ... QMNGK ... NPTPQ ... CFNHN ... NERDN ... QYVYN ... FRRFQ ... HTGNT ... KYECR ... ... TGQWK ... CDKRF ... QMNGK ... NPMPQ ... CFNHN ... NERDN ... QYVYN ... FRRFQ ... HTGNT ... KYECR ... ... TGQWK ... CDKRF ... QMNGK ... NPTPQ ... CFNHN ... NERDN ... QYVYN ... FRQFQ ... HTGNT ... KYECR ... ... TGQWK ... CDKRF ... QMNGK ... NPPPQ ... CFNHN ... NERDN ... RYVYN ... FRRFQ ... HTGNT ... KYECR ... p.Q152R
p.K430K
p.N470N
p.T523A p.N601S p.R810Q p.V816I p.R959Q p.G1084E p.E1166Q
(a) Domain architecture of human thrombospondin 1 (TSP1), and variants identified in different domains. This architecture was drawn according to the data on human TSP1 (P07996) in the website http://web.expasy.org/docs/swiss-prot_guideline.html. Locations of the variants in the deferent domain are shown with black arrows. (b) Alignment of the TSP1 protein sequence with homologous protein sequences of other species. Most variants were completely conserved in different species. Amino acids that were not conserved are shaded in gray. EGF, epidermal growth factor-like domains; LG, laminin G-like amino-terminal domain; L-lectin, L-type lectin-like domain; TSR, thrombospondin type 1 domains; Type 3, thrombospondin type 3 repeats; vWF_C, von Willebrand factor type C domain.
Table 5
Functional prediction and conservation analysis of nonsynonymous variants using SIFT, PolyPhen2, and GERP + +
Variant c.455A > G:p.Q152R c.1567A > G:p.T523A c.1802A > G:p.N601S c.2429G > A:p.R810Q c.2446G > A:p.V816I c.2876G > A:p.R959Q c.3251G > A:p.G1084E c.3496G > C:p.E1166Q
dbSNP
SIFT_score
SIFT_pred
PolyPhen2_score
PolyPhen2_pred
GERP + +
rs183783284 rs2292305 rs201454661 NA rs143790069 NA NA NA
0.97 0.18 0.17 0.69 0.92 0.39 1 0.87
D T T T T T D T
0.028 0 0 0.179 0.041 0.334 1 0.988
B B B P B P D D
4.53 −1.26 4.2 5.35 5.35 4.34 5.43 5.39
B, benign; D, damaging; dbSNP, single nucleotide polymorphism database; P, possibly damaging, T, tolerant.
Acknowledgements The authors thank all the families that participated and collaborated with us in this study. They also thank the Yilin (Elim) Autism Training Department of the Qingdao Municipal Autism Research Institute, which provided a lot of help in the collecting samples. They also thank Professor Jiada Li, who assisted with proofreading of the manuscript. This work was supported by the National Basic Research Program of China (2012CB517902) and
the National Natural Science Foundation of China (81330027, 81161120544 and 81301172). Correspondence should be addressed to Kun Xia ([email protected]) or Zhengmao Hu (huzhengmao@ sklmg.edu.cn). Conflicts of interest
There are no conflicts of interest.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
240 Psychiatric Genetics 2014, Vol 24 No 6
References Aberle H, Haghighi AP, Fetter RD, McCabe BD, Magalhães TR, Goodman CS (2002). Wishful thinking encodes a BMP type II receptor that regulates synaptic growth in Drosophila. Neuron 33:545–558. Abrahams BS, Geschwind DH (2008). Advances in autism genetics: on the threshold of a new neurobiology. Nat Rev Genet 9:341–355. Adams JC, Lawler J (2011). The thrombospondins. Cold Spring Harb Perspect Biol 3:a009712. Adams JC, Bentley AA, Kvansakul M, Hatherley D, Hohenester E (2008). Extracellular matrix retention of thrombospondin 1 is controlled by its conserved C-terminal region. J Cell Sci 121:784–795. Barrett JC, Fry B, Maller J, Daly MJ (2005). Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics 21:263–265. Beaudet AL (2002). Is medical genetics neglecting epigenetics? Genet Med 4:399–402. Christopherson KS, Ullian EM, Stokes CC, Mullowney CE, Hell JW, Agah A, et al. (2005). Thrombospondins are astrocyte-secreted proteins that promote CNS synaptogenesis. Cell 120:421–433. Dale O, Salo M (1996). The Helsinki Declaration, research guidelines and regulations: present and future editorial aspects. Acta Anaesthesiol Scand 40:771–772. Eroglu C, Allen NJ, Susman MW, O’Rourke NA, Park CY, Ozkan E, et al. (2009). Gabapentin receptor alpha2delta-1 is a neuronal thrombospondin receptor responsible for excitatory CNS synaptogenesis. Cell 139:380–392. Graf ER, Zhang X, Jin SX, Linhoff MW, Craig AM (2004). Neurexins induce differentiation of GABA and glutamate postsynaptic specializations via neuroligins. Cell 119:1013–1026. Lawler J, Simons ER (1983). Cooperative binding of calcium to thrombospondin. The effect of calcium on the circular dichroism and limited tryptic digestion of thrombospondin. J Biol Chem 258:12098–12101. Lawler J, Hynes RO (1986). The structure of human thrombospondin, an adhesive glycoprotein with multiple calcium-binding sites and homologies with several different proteins. J Cell Biol 103:1635–1648. Lorentzon M, Hoffer B, Ebendal T, Olson L, Tomac A (1996). Habrec1, a novel serine/threonine kinase TGF-beta type I-like receptor, has a specific cellular expression suggesting function in the developing organism and adult brain. Exp Neurol 142:351–360.
Marqués G, Bao H, Haerry TE, Shimell MJ, Duchek P, Zhang B, O’Connor MB (2002). The Drosophila BMP type II receptor Wishful Thinking regulates neuromuscular synapse morphology and function. Neuron 33: 529–543. Mehler MF, Mabie PC, Zhang D, Kessler JA (1997). Bone morphogenetic proteins in the nervous system. Trends Neurosci 20:309–317. Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D, et al. (2007). PLINK: a tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet 81:559–575. Schultz-Cherry S, Lawler J, Murphy-Ullrich JE (1994). The type 1 repeats of thrombospondin 1 activate latent transforming growth factor-beta. J Biol Chem 269:26783–26788. Schuman EM (1997). Synapse specificity and long-term information storage. Neuron 18:339–342. Südhof TC (2008). Neuroligins and neurexins link synaptic function to cognitive disease. Nature 455:903–911. Sun X, Skorstengaard K, Mosher DF (1992). Disulfides modulate RGD-inhibitable cell adhesive activity of thrombospondin. J Cell Biol 118:693–701. Varoqueaux F, Aramuni G, Rawson RL, Mohrmann R, Missler M, Gottmann K, et al. (2006). Neuroligins determine synapse maturation and function. Neuron 51:741–754. Wang K, Li M, Hakonarson H (2010). ANNOVAR: functional annotation of genetic variants from high-throughput sequencing data. Nucleic Acids Res 38:e164. Washbourne P, Dityatev A, Scheiffele P, Biederer T, Weiner JA, Christopherson KS, El-Husseini A (2004). Cell adhesion molecules in synapse formation. J Neurosci 24:9244–9249. Waterhouse L, Morris R, Allen D, Dunn M, Fein D, Feinstein C, et al. (1996). Diagnosis and classification in autism. J Autism Dev Disord 26:59–86. Withers GS, Higgins D, Charette M, Banker G (2000). Bone morphogenetic protein-7 enhances dendritic growth and receptivity to innervation in cultured hippocampal neurons. Eur J Neurosci 12:106–116. Xu J, Xiao N, Xia J (2010). Thrombospondin 1 accelerates synaptogenesis in hippocampal neurons through neuroligin 1. Nat Neurosci 13:22–24. Zhang F, Endo S, Cleary LJ, Eskin A, Byrne JH (1997). Role of transforming growth factor-beta in long-term synaptic facilitation in Aplysia. Science 275:1318–1320.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
